An Experimental Study of Adaptive Application-Sensitive Partitioning Strategies for SAMR Applications

Dynamic adaptive mesh renement methods for the numerical solution to partial dierential equations yield highly advantageous ratios for cost/accuracy when compared to methods based upon static uniform approximations. Parallel/distributed implementations of these techniques, however, present signicant challenges in dynamic data-distribution and load-balancing. This is because the choice of the \best " partitioning technique and associated partitioning parame-ters depends on the nature of the application and its run-time state. This paper presents an experimental study of an adaptive application-sensitive meta-partitioner for structured adaptive mesh renement (SAMR) applications that dynamically selects and congures partitioning strate-gies at run-time based on system parameters and the current application state. The selection is based on a run-time classication of application state and an application-centric classication of the partitioners. Experimental results show that adaptive partitioning can improve application performance